Direct answer
Yes, but rarely on the broad question. In a maxaeo panel of 1,248 B2B software prompts across eight AI engines, smaller challenger brands got 14% of recommendations on generic "best [category]" prompts and 38% once the prompt named a use case. So stop chasing "best CRM" and own the specific questions your buyers ask. Named tracks those per engine, on a fixed prompt set it repeats.
Do AI engines favor big brands?
On broad questions, yes. When the prompt is just a category name, engines reach for the names they can most easily back up, and that's usually the incumbents.
The cleanest test is a June 2026 preprint by Xi Chu and Yupeng Hou. They gave GPT-4o-mini, Claude Sonnet and Gemini 3 Flash ten skincare products with identical specs: one real brand and nine made-up ones. The real brand got picked every time. But give an unknown product a rating edge of less than a tenth of a star and that lead mostly disappeared. Brand works like a tie-breaker. It decides things when nothing else does.
maxaeo found the same shape in software. Company size, age and funding explained 22% of the variation in how often a brand got mentioned. Adding third-party coverage and use-case proof pages raised that to 41%.
Which questions can a small business win?
The narrow ones, where the buyer adds a constraint that a big brand's generic pages don't speak to.
| Prompt type (maxaeo panel, May to June 2026) | Challenger share of recommendations |
|---|---|
| Generic "best [category]" | 14% |
| Integration ("works with X") | 34% |
| Compliance and security | 35% |
| Migration ("moving off X") | 37% |
| Use case ("for [specific job]") | 38% |
| Startup or mid-market ("for a team of 10") | 41% |
These are vendor numbers from B2B SaaS and tech prompts only. maxaeo says local services and ecommerce may look different. I still think the direction holds for most small businesses: "best plumber" goes to whoever has the most reviews everywhere, but "plumber who does tankless water heater installs in North Brunswick" has far fewer contenders.
Why does a specific page beat a famous name?
Because the engine needs something to quote that answers the exact question. A big brand's homepage says "for teams of every size." A smaller company's page that says "built for 5-person dental offices, works with Dentrix, here's what it can't do" gives the engine a direct match.
maxaeo calls this evidence matching, and says the vendors that won narrow prompts had migration guides, integration docs, security pages, comparison pages and case studies for that exact use case. Every item on that list is a page a real buyer could use.
Does third-party coverage matter more than my own site?
For being remembered across engines, it looks that way. Fractl ran 96 industry prompts 15 times each on GPT-4o, Gemini 2.5 Flash and Claude Sonnet 4.6. The brands that showed up more than their search authority predicted had one thing in common: they kept appearing in other people's roundups, comparison articles and category coverage.
The same study found only 11% of brands were named by all three models, and 77% by just one. So a small business can be strong in Claude and missing in Gemini at the same time. One blended "AI visibility score" hides that.
What shouldn't a small business do?
Don't fake proof. The Chu and Hou preprint found that made-up clinical claims could break a big brand's lead in their tests. Read that as a warning about the channel. Fake claims can get a business in trouble with customers and regulators, and Google's AI optimization guide warns against pages made mainly to manipulate AI answers.
Also skip these:
- Your own "10 best [category] tools" list with you at #1. It hands the engine a ready-made list of your competitors.
- One thin page for every possible question. Coverage without anything new to say doesn't help.
- Judging yourself on the broadest prompt in month one. That's where incumbents are strongest.
How Named tracks this
Named keeps one fixed prompt set and repeats it to check changes. The Namies draft fixes, and a person checks every change. Results vary run to run, so Named repeats each question before calling a change, and says when the sample is too small.
| What Named does | Why it matters for a small business |
|---|---|
| Keeps a fixed list of your buyers' prompts | Put the narrow questions on it, since those are the ones you can win |
| Counts each engine separately | You can be named by one engine and missing from another |
| Logs who gets named instead of you | Shows which competitor owns each question |
| Logs the links each answer cites | Shows which pages, yours or someone else's, the engine read |
Named checks ChatGPT, Claude, Gemini, Perplexity, Grok and Muse (muse.ai). The rules are on our facts page. For why being named and being linked are different counts, see Mentioned vs cited, and for why traffic is the wrong yardstick, see AI visibility vs AI traffic.
FAQ
Can a brand-new business get recommended by ChatGPT?
Sometimes, on narrow questions where it has a page that answers that exact need and some coverage elsewhere. It's unlikely on the broad category question. Nobody can promise it, and answers change from run to run.
Is it worth trying to rank for "best [category]" at all?
Track it, but don't start there. In maxaeo's panel that's where challengers got the smallest share.
Do reviews and directory listings help?
Fractl's overperformers were the brands that kept appearing in third-party content. That's correlation, not proof, but it's the most consistent pattern across the studies here.
Which engine should a small business check first?
All the ones your buyers use, each counted on its own. Fractl found most brands show up in only one of the three models it tested. You can start with the free audit.
Sources
- maxaeo (Chris Han), "Does ChatGPT Favor Big Brands? Evidence and 2026 Tests," Jul 8, 2026: https://maxaeo.ai/blog/does-chatgpt-favor-big-brands/ (vendor field study, 1,248 prompts, 52 B2B SaaS and tech categories, 8 engines, 139,776 answer observations; read Oct 27, 2026)
- Xi Chu and Yupeng Hou, "Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems," arXiv:2606.17443, v2 Aug 21, 2026: https://arxiv.org/abs/2606.17443 (academic preprint, not peer reviewed; read Oct 27, 2026)
- Fractl, "Which Major Brands Are Vanishing From ChatGPT, Gemini, and Claude": https://www.frac.tl/ai-visibility-index/ and Search Engine Land write-up https://searchengineland.com/ai-visibility-index-brands-vanishing-from-ai-search-485057 (vendor study; read Oct 27, 2026)
- Google AI optimization guide, updated Jul 10, 2026: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide (Google policy; read Oct 27, 2026)
- Named facts page: https://named.khanex.ai/facts/ (read Oct 27, 2026)
About the author: Saad Khan is a 23-year-old AI engineer and the founder of Named, which works to get small businesses recommended by AI engines.